The overall goal of the "mitogenic signaling networks" program is to develop high level statistical and specific physico-chemical models that described key features of mitogenic signaling networks activated by ErbB receptors (EGF receptor, and Neu2/Her2) and by oncogenic K-ras. We will use numerical models developed with cells in culture to explore tumor formation in mice genetically engineered to recapitulate key features of human disease. Ultimately, it is our hope that numerical models will predict the responsiveness of various tumors to anti-neoplastic therapies that target ras and ErbB. Currently these therapies work in only a subset of patients and must therefore be used in combination with, or after, conventional cytotoxic therapy. Throughout this project we will integrate data on immediate early responses obtained from biochemical analysis of receptors, adapter proteins, kinases and small GTPases with subsequent transcriptional responses and the acquisition of transformed phenotypes. This integration will be achieved through models designed to capture as much biological prior knowledge and experimental data as possible. In the initial phases of our work, physico-chemical models will be used when the amount of data is high and the biological complexity relatively low (cells grown in culture, for example) whereas statistical modeling will be used when the problems are less well defined and more complex. However, the future lies with hybrid models that link detailed mechanistic data from cells grown in culture to data on the behavior of cells in actual tumors.